evaluation of water produced from a steam dominated … · evaluation of water produced from a...
TRANSCRIPT
Proceedings World Geothermal Congress 2015
Melbourne, Australia, 19-25 April 2015
1
Evaluation of Water Produced from a Steam Dominated System, a Case Study from the
Darajat Field
Christovik H. Simatupang1, Rindu G. Intani
1, M. Ramos Suryanta
1, Riki Irfan
1, Glenn Golla
1, Cedric Cease
1 and Phil
Molling2
1Chevron Geothermal Indonesia, Sentral Senayan Office Tower II, 26th Floor, Jl. Asia Afrika No. 8 Jakarta 10270, Indonesia
2G7geochem, Santa Rosa, CA USA
[email protected], [email protected], [email protected], [email protected],
[email protected], [email protected], [email protected]
Keywords: Water chemistry, steam dominated system, marginal recharge, injection
ABSTRACT
Darajat is unusual as it is one of the few steam-dominated geothermal systems in the world. Current thinking indicates that the
reservoir brine in this geothermal system is deeply located and has not been accessed yet, and therefore, liquid is unlikely to be
produced from the existing production wells. DRJ-CA, drilled in early 2008 in the central part of the field, is an enigma because it
produces liquid.
Pressure-Temperature (PT), Pressure-Temperature-Spinner (PTS), and Tracer Flow Test (TFT) data showed that a maximum of
~74 kg/s of liquid with an enthalpy of 916 BTU/lb is being produced from this well both through shallow and deep entries. Two
permeable entries just below the production casing are the suspected sources of the liquid but the exact location liquid entry is quite
difficult to delineate because of the nature of these two feed zones. The deep liquid feed zone is at saturation temperature and is
believed to flash to steam and very unlikely to contribute to liquid at the surface. To prevent liquid to be carried over into the steam
pipeline, DRJ-CA is operated at high Flowing Well Head Pressure (FWHP).
Geochemical analysis of Down Hole Samples (DHS) indicates that the liquid encountered by DRJ-CA is typical of steam-heated
groundwater above or outside the reservoir with significant silica content. This suggests that there may be Marginal Recharge (MR)
down flow in this part of the field. A complication with this hypothesis is that power plant condensate has been injected in both
shallow and deep wells right at the center of the field. Three hypotheses are postulated to explain the source of the liquid in DRJ-
CA, namely injected condensate, MR entry into the reservoir, or a combination of both. The recent strategy to move infield
condensate injection to the northeastern edge of the field should help further narrow the source of the DRJ-CA liquid.
1. INTRODUCTION
The Darajat geothermal field is located about 150 kilometers southeast of Jakarta in West Java, Indonesia at an elevation of 1,750–
2,000 m Above Sea Level (ASL) (Figure 1). It is situated on the eastern side of Mt. Kendang which is part of a NE–SW-trending
Quaternary volcanic range in West Java. There are numerous eruptive centers within this mountain range and numerous volcanic
activities occurred, for example, at the Papandayan Volcano in 1772 and in the Guntur Volcano in 1840.
Darajat is one of the largest vapor-dominated geothermal fields in the world with a current total capacity of 271 MWe,
accommodated by three power plants (Units I/II/III). PT. Indonesia Power owns and operates Unit I with a generation of 56 MWe,
while Units II/III (with 94 MWe and 121 MWe of generation, respectively), are operated by Chevron Geothermal Indonesia.
Figure 1: Location of the Darajat Geothermal Field in relation to other cities in the Java Island.
The most significant structural feature in this area is the Kendang Fault which strikes northeast from Darajat along the high axis of
the volcanic range, disappearing on the north side of the Kamojang field, another steam-dominated geothermal system about 10 km
Simatupang et al.
2
away. To the west of the Darajat field, the Kendang Fault is slightly offset by the Gagak Fault, which is considered to control major
permeability within the field (Figure 2).
Figure 2: Map showing the Darajat geothermal field. Also shown are the locations of DRJ-CA and other wells mentioned in
this paper. The red shaded area is the interpreted extent of the shallow perched aquifer, while the numbers show the
temperature of fluid in the shallow perched aquifer measured in wells.
Since the first investigation in the early 1970’s, Chevron Geothermal Indonesia has drilled 49 wells in seven development and
make-up drilling campaigns; the last make-up drilling campaign was during 2009–2011. These include 43 development,
production, and injection and six slimhole observation wells. Historically, four wells have produced liquid since the first test in the
late 1970s, namely DRJ-B, DRJ-C, DRJ-E, and the recent DRJ-CA (Figure 2).
During the development phase of the Darajat field, DRJ-B and E were utilized as observation wells, while DRJ-C was used for
injecting the power plant condensate from Unit I. DRJ-CA, the main subject of this paper, was drilled in February 2008 along a
NW direction at a moderate inclination. It was designed to produce steam from the good permeability region between the Gagak
and Cibeureum Faults. The top of the reservoir encountered by DRJ-CA was defined at about 720 m ASL; this is relatively shallow
compared with other neighboring wells. Based on the injection testing after well completion, DRJ-CA was estimated to produce
about 17 kg/s (~8 MWe) of steam at 16 bara Flowing Well Head Pressure (FWHP).
Right after completion, the DRJ-CA heat-up data showed temperature reversals at, from top to bottom, 1,700 m ASL, 1,500 m
ASL, 1,400 m ASL, 400 m ASL, and 50 m Below Sea Level (BSL) (Figure 3). The shallower temperature reversals above 1,400 m
ASL were interpreted to be related with the shallow aquifer recently mapped in Darajat. The reversals at about 400 m ASL, near the
production casing shoe, and at 50 m BSL were suspected to be related with cold fluid influx. Liquid production was first observed
during the flow test in July 2008 and a later Tracer Flow Test (TFT) showed 1.5–9.3 kg/s of liquid was being produced from this
well at 18–21 bara FWHP. DRJ-CA was eventually flowed to the system with 5 kg/s of steam at 23 bara FWHP; at this higher
FWHP, the well produces dry steam.
The discovery of produced liquid at DRJ-CA was unexpected. Current thinking indicates that the reservoir brine in the Darajat
geothermal system is located deep and has not been accessed by any wells. Therefore, it is unlikely to be produced from the
existing production wells. Possible sources of this liquid may be the infield condensate injected at injectors DRJ-AE and C, entry of
Marginal Recharge (MR) into the reservoir, or a combination of both. In addition, some wells, such as DRJ-B, C, and E, produced
liquid in the 1980s, while there was no infield injection yet suggesting that MR appears to be the source of liquid at Darajat. This
paper further discusses the reservoir processes that may be responsible for the presence of liquid in Darajat.
AE
C
B
E
CA
Unit II & III 215 MWe
Unit III56 MWe
150°
130°
56°
50°
56°
81°
95°
160°
47°
80°
Extent of Aquifer
150° Aquifer temp (°C)
Aquifer flow
Fumarole area
WellsLiquid WellsInjection Wells
B0
Simatupang et al.
3
Figure 3: Chart showing heat-up PT data of DRJ-CA. Note the temperature reversals at about 1,400 m ASL, 400 m ASL,
and 100 m BSL. Also shown are the kaolinite and argillic alteration zones (blue-shaded region) encountered by the
well.
2. TAGGING THE PRODUCED LIQUID AT DRJ-CA
2.1. Pressure-Temperature-Spinner (PTS) Analysis
To get a better understanding of well deliverability and reservoir characterization, an initial series of well testing program was
conducted at DRJ-CA. PTS analysis in 2009 indicated liquid feed zones near the production casing shoe and near TD (Figure 4).
Most of the mass produced by DRJ-CA comes from the permeable steam entries at 3,500’–4,000’ MD. However, the exact location
of where the liquid enters near these two feed zones could not be clearly defined from the log data. Currently, both feed zones are
considered as liquid entries (Figure 4). The liquid produced by the bottom permeable entry at 6,500’ MD is at saturation and does
not contribute significantly to the total well deliverability. This liquid probably flashes to steam as it enters the wellbore and is
unlikely to contribute much liquid (and mass) at the surface.
Figure 4: Chart showing the permeable entries encountered at DRJ-CA during an injection/flowing PTS. The feed zones at
3,500’–4,000’ MD are suspected to produce liquid although the log data cannot distinguish which particular feed
zone the liquid is coming from. The produced liquid from the bottom permeable entry probably flashes and does not
contribute significantly to total well deliverability at the surface.
Deep liquid sourceShallow liquid source
DRJ-CA PTS
production casing shoe
Simatupang et al.
4
In terms of productivity, DRJ-CA is not a good production well. This well exhibited slow recovery after a Minimum Quenching
Rate Test, unlike other Darajat steam producers. The slow heat-up indicates that this well reached saturation after 36 days of
injection and was ultimately confirmed by the production test which showed that this well has an initial production of about 17 kg/s
at 16 bara. This production rate is significantly less compared with the other Darajat wells drilled during that time. DRJ-CA was
first put on production almost a year after its completion in December 2008 and produced at about 5 kg/s at 23 bara FWHP. This
pressure was considered as the minimum FWHP to avoid liquid production from this well.
2.2. Hydrothermal Alteration
Analyses of hydrothermal alteration, temperature anomalies, and the location of shallow permeability have delineated a shallow
aquifer at Darajat (Rohrs et al., 2009). This aquifer is believed to exist in the southeastern portion of the field and probably flows
northeast towards the Cibeureum and Toblong hot spring areas, passing through the DRJ-B location (Figure 2).
Reservoir steam at the fumarolic areas produces acidic fluids when the fumarolic steam condenses and mixes with local ground
water. The ground water possibly becomes acidic as H2S (entrained in the steam) is oxidized to form acid-sulfate fluid; further
fluid-rock interaction leads to the formation of secondary hydrothermal alteration minerals, such as kaolinite, alunite, etc. At
Darajat, the perched steam condensate may have been “acidified” with the continuous introduction of H2S entrained in the steam.
Although this process commonly occurs in a geothermal system, the acid nature of this perched aquifer is still questionable because
no sample has been collected at the suspected depth of the shallow perched aquifer.
Post-drilling data show that DRJ-CA demonstrated some evidence for the existence of a shallow perched aquifer as characterized
by temperature reversals and secondary kaolinite at 1,545–1,114 m ASL (Figure 3). However, this shallow perched aquifer
encountered by DRJ-CA is relatively deeper compared with its neighboring wells (Figure 5).
Figure 5: 3D cross-section showing the shallow aquifer encountered by DRJ-CA and other neighboring wells. Note that the
shallow aquifer deepens towards the north as intersected in DRJ-CA. The color represents the temperature of the
shallow aquifer as measured in wells.
2.3. The Clay Cap Above the Reservoir
Typical of other geothermal systems, the rocks that cap the Darajat geothermal system are altered to secondary clays. These clays
vary from smectite (stable at ≤180ºC), mixed-layered illite-smectite or smectite-chlorite, etc. A re-interpretation of the Magneto-
Telluric (MT) data at Darajat using the 1D invariant mode defined the location of the 10 ohm-m clay-altered rocks that overlie the
geothermal system (Figure 6). The distribution of the low resistivity (Figure 6) and mineral alteration (Figure 3) indicates that the
clay cap distribution is consistent. The perched aquifer appears to be located within the clay cap.
Figure 6: Cross section A-A’ showing the clay cap (green shaded area) distribution atop the Darajat reservoir.
Simatupang et al.
5
2.4. Fluid chemistry
The chemistries of both produced liquid and down hole liquid samples at DRJ-CA are very distinctive when compared with the
power plant condensate injectate, steam condensate collected at drain pots (DPs) and other low point drains (LPDs) along the steam
pipelines, and surface water (Table 1). However, the DRJ-CA produced liquid is similar to the Awibengkok shallow leak liquid.
Condensate injectate can be distinguished from other fluids by its neutral pH, and low SiO2, low HCO3, and relatively higher SO4
contents. Liquid from DPs and LPDs show extremely high B with 5–6 pH; the high B content can be explained by continuous
boiling from flowing steam to static condensate liquid which accumulates at the bottom of the pipeline. Surface water, which does
not undergo significant water-rock interaction, has low SiO2, B, SO4 and HCO3 contents with neutral pH. In the geothermal
reservoir, the produced liquid is characterized by high SiO2, high SO4, high HCO3, and relatively high B contents and normal to
basic pH values like the Awibengkok shallow leak liquid (Table 1).
Table 1: Chemistries of possible Darajat liquid source compared with the Awibengkok shallow leak liquid
If the produced liquid at DRJ-CA is from the injected power plant condensate, the condensate must have reacted with the reservoir
rocks to acquire elevated concentrations of Na, K, Ca, SiO2, and Cl. This reaction is further complicated by the fact that
condensation and mixing with the fluid that has reacted with the rocks should also generate large amounts of dissolved B, HCO3,
and SO4. The reactive history of the condensate injectate metamorphosing into the produced liquid at DRJ-CA would require
opposing reservoir processes, i.e., mixing of two liquids and equally important constant exposure to “fresh” reservoir rocks that are
not coated by reaction products of the liquid–rock interaction.
Three reservoir processes are required to explain the chemical conversion from condensate-type to a liquid similar to that of DRJ-
CA, and these processes are as follows:
NCG influx: significant increase in dissolved SO4 and HCO3 requires high flux of NCG through the liquid (e.g., bubbling
and condensation);
Sustained boiling: significant increase in dissolved B requires sustained boiling/evaporation of liquid to concentrate B in
liquid (e.g., steam/liquid fractionation); and
Water-rock reaction: significant increase in SiO2 requires sustained reaction of dilute liquid with rock/clays (e.g., enhanced
by low pH and/or high temperature).
The produced liquid and DHS chemistry of DRJ-CA plot in the steam-heated water portion of the Cl-SO4-HCO3 ternary diagram
(Figure 7). The chemistry of this liquid was also compared with the Awibengkok shallow leak liquid sample (a possible perched
aquifer liquid at Awibengkok) and condensate injectate liquid from the Darajat injectors. All of the condensate chemistry plots at
the SO4 apex. The condensate injectate is depleted in HCO3 compared with the produced liquid from DRJ-CA. The presence of
chloride and other chemical species in the DRJ-CA liquid indicate that this liquid is not directly related to the injected power plant
condensate.
Liquid TypeBoron
(mg/kg)SiO2
(mg/kg)SO4
(mg/kg)HCO3
(mg/kg)Cl
(mg/kg)pH
Condensate Injectate 15 – 57 0.5 – 2 45 – 90 < 16 0 6 – 7
Condensate Pot 60 – 1150 0.5 – 32 0.2 - 31 2 – 5 0 5.5 – 6.5
Low Point Drain 1200 – 1500 20 – 50 2 – 5 6 – 6.5 0 5 – 6
Surface Water < 1 14 – 17 3.5 – 4 9 – 12 0 7
Awibengkok Shallow Leak Liquid
9 168 550 103 186 6
DRJ-CA DHS 164 505 340 102 6 – 200 7
DRJ-CA Liquid (Surface Sampling)
3 – 250 150 - 600 150 – 600 150 – 400 30 – 65 6 – 9
DRJ-C, DRJ-B, DRJ-E 5 – 100 400 – 1000 150 – 500 90 – 250 30 – 150 7 – 9
Simatupang et al.
6
Figure 7: Ternary plot of Cl-SO4-HCO3 showing the liquid chemistry from surface discharge and DHS at DRJ-CA. Red
circles represent surface samples while green triangles are DHS. Also shown is the Salak shallow leak liquid for
comparison. The liquid chemistry matches with steam-heated water characteristics.
3. RESERVOIR MODELS TO EXPLAIN THE LIQUID AT DRJ-CA
As stated earlier, NCG influx, sustained boiling and water-rock reaction need to be incorporated into the models to explain
chemical conversion of a dilute liquid to something similar to DRJ-CA liquid. Three models are proposed with each model
requiring different well conditions, various liquid source/s, and different reservoir processes as described below.
3.1. Model 1: Boiling of condensate injectate
Model 1a - This end-member DHS liquid chemistry is characterized by high concentrations of SiO2, HCO3, and SO4. Assuming
condensate injectate is the source of the DHS and the well is in flowing condition, the sequence of the reservoir processes that
should occur is (1) water-rock/clay reaction, (2) condensation, and (3) boiling which occurs along the fractures during fluid
migration. This model is applicable to DRJ-CA discharge liquid and other flowing scenarios. The downside of this model is the
continuous source of fresh rock that will be exposed to liquid to maintain the high SiO2 content (i.e., if the same rock is exposed to
interaction with flowing fluids, the silica content of the rock will have been consumed after some period).
Model 1b – Similar with the above model, this end-member DHS liquid chemistry contains high SiO2, HCO3, and SO4
concentrations. Assuming condensate injectate is the source of the liquid and the well is in flowing condition, the order of the
reservoir processes that should happen is (1) condensation together with (2) water-rock/clay reaction along the fractures during
fluid migration; (3) boiling of the liquid occurs at the bottom of the wellbore to enrich the SiO2 and B. This model only applies to
shut-in conditions, so liquid fills up the bottom of the well (or wells) which are close to the injectors (i.e., within the “cone of liquid
migration”). If a well is located far away from injector, the source of the liquid should be replaced with a natural steam condensate.
The downsides of this model are similar with the Model 1a above.
Figure 8: Model 1a – produced liquid has undergone continuous boiling and water-rock reaction along fracture pathways.
Model 1a Cartoon (DRJ-31 Flowing Liquid)
condensateinjectate
SiO2
B
SO4
HCO3
condensation
boiling
Model 1a: Schematic of Continuous Boiling of Injected Condensate
Simatupang et al.
7
Figure 9: Model 1b – produced liquid underwent water-rock reaction along fracture pathways and boiling occurs at the
bottom of the wellbore during shut-in conditions.
3.2. Model 2: Condensate injectate undergoes similar conditions as Condensate/Low Drain Pots
This model assumes that boiled condensate injectate (i.e., high B liquid) is the source of the liquid and then undergoes a CP- and
LPD-like process to explain the chemical change from dilute fluid to DRJ-CA liquid type. The CP- and LPD-like process provides
continuous boiling along the fractures during fluid migration to wellbore. The sequence of reservoir processes should be (1) water-
rock/clay reaction in order to enrich the SiO2 content, (2) condensation may happen next creating a high flux of NCG passing
vertically through the fractures where liquid travels enriching both the SO4 and HCO3, and (3) the last process should be boiling to
ensure the SiO2 and B content remains high per DHS liquid results. This model appears to be limited as continued fresh rock
exposure is needed to maintain the SiO2 content (please refer to downsides of both Models 1a and 1b). Also, constant fracturing or
re-routing of the liquid to fresh fractures would be required to allow large-scale production of this fluid type.
Figure 10: Model 2 –liquid has undergone boiling, condensation, and further boiling along fracture pathways as it travels
into the wellbore (i.e., CP- and LPD-like process).
3.3. Model 3: MR liquid undergoes enrichment from rising NCG and downflows to the geothermal reservoir
In this model, shallow MR liquid (which may be above or within the clay cap) may experience continuous condensation as NCG
passes through it. Fracture permeability allows fluid from the reservoir to escape vertically and arrive near the surface. The
exposure of groundwater to reservoir NCG increases both the HCO3 and SO4 concentrations. As the pressure drops within the
geothermal reservoir due to mass extraction, this MR liquid may enter the reservoir along the same fluid pathways originally used
to “steam-heat” the groundwater. The pressure drop caused by the production wells causes the high SO4 and HCO3 liquid to
migrate (possibly laterally) and travel down to the wellbore and intersected permeable production entries. During the lateral
migration, water-rock/clay reaction and, possibly additional boiling increases both SiO2 and B concentrations. These processes
provide chemical conversion from dilute to a liquid similar to that produced at DRJ-CA.
Model 2 Cartoon
B
SiO2
Boiled condensate injectate
SO4
HCO3
condensation
boiling
Model 1b: Schematic of Boiling of Injected Condensate in the Wellbore
Model 1a: Schematic of Continuous Boiling of Injected Condensate
Simatupang et al.
8
Figure 11: Model 3 - MR liquid undergoes enrichment from rising NCG and downflows into the geothermal reservoir along
permeable fractures.
Among these three reservoir models, Model 3 (Figure 11) is preferred because of the following reasons:
This model does not need continuous exposure to fresh rock to maintain the high SiO2 contents in flowing shallow liquids, as
a small reservoir of liquid is created above the clay cap (i.e., at DRJ-CA);
Geologic time allows this complicated conversion processes to occur outside of the reservoir;
Chemical analogs of shallow “steam-heated” liquids exist (e.g., Awibengkok shallow leak liquid sample); and
Other models are too complex for “real-time” chemical conversion (“Occam’s Razor Principle”).
In 2012, infield injection in DRJ-AE was moved to the edge of the field and has helped to confirm whether MR liquid or
condensate injectate supplies the liquid to the production wells. A liquid similar to the previous chemistries is still being produced
at DRJ-CA even after infield injection has been terminated. Therefore, a neutral peripheral/marginal liquid (e.g., perched aquifer) is
believed to represent the most likely model (Model 3) to explain the presence of the liquid produced in DRJ-CA
4. CONCLUSIONS
The multidisciplinary approach in studying the liquid produced at the steam-dominated Darajat geothermal field has assisted in
defining the location of the source of the liquid produced at DRJ-CA and the possible reservoir processes responsible for this liquid.
The relatively high concentrations of SiO2, HCO3, SO4, B, and Cl and neutral pH of the produced liquid suggests that this liquid has
had sufficient time to react with reservoir rocks and is probably not the injected power plant condensate. In addition, the presence
of both HCO3 and SO4 indicate that it may be peripheral or near-surface liquid. It is believed that this liquid comes from a shallow
aquifer downflowing into the Darajat reservoir.
To produce dry steam, DRJ-CA is flowed at ~16 barg FWHP. The plan forward is to put this shallow aquifer behind casing in
future make-up wells, and evaluate isolating this aquifer in the existing wells.
ACKNOWLEDGMENT
We thank Chevron Geothermal Indonesia (CGI) for supporting this project and Pertamina Geothermal Energy for granting
permission to publish this paper.
REFERENCES
Rohrs, D., Rejeki, S., Nordquist, G., Molling, P., Sugandhi, A., Acuna, J., Pasaribu, F., Fitriyanto, A. and Kurniawan, A.: Darajat
2007-08 Static Model Update: CGI Internal Report, (March 30, 2009).
Simatupang, C. and Molling, P.: The Case for an External Origin for MR - based on Liquid Chemistry, CGI Internal Presentation,
(August, 2010), 45 slides.
Simatupang, C. and Molling, P.: 2009 Darajat Geochemistry Annual Report: CGI Internal Report, (November 5, 2010), 38 pages.
Simatupang, C. Molling, P. and Sunio, E.: The Mid-2009 Geochemistry Monitoring Report: CGI Internal Report, (October, 13,
2009), 23 pages.
Simatupang, C., Mahagyo, P. and Molling, P.: The Geochemistry of Produced Liquid at DRJ-CA: CGI Internal Presentation,
August 9, 2009, 16 slides.
Ussher, G., Harvey, C., Johnstone, R., and Anderson, E.: Understanding the Resistivities Observed in Geothermal System,
Proceedings, World Geothermal Congress 2000, Kyushu – Tohoku, Japan, 2000, 6 pages.
Condensation
Model 3 Cartoon
Surface Manifestation
Clay Cap
Steam
Neutralization
Clay Cap
BSiO2 SO4
HCO3